4 Mr. D. Thoday. Experimental Researches on [June 11, 



of the increase is briefly dealt with in Section III ; the others are outside 

 the scope of the present paper : — 



Summary of Sources of Error affecting the Half-leaf Dry Weight Method. 



i. Errors inti'oduced in any single case in estimating the dry weight 



of the unit of area of leaf lamina : 



a. In measuring or determining the ai'ea, by the various methods Section VI. 



b. In determining dry weight : 



(1) . In drying Section VII. 



(2) In weighing. 



ii. Errors introduced in the comparison of two halves of a leaf : 



a. Through asymmetry due to : 



(1) Un.symmetrical venation ; "I Section IV 



(2) Inequalities in thickness ; J " " ^ 



(3) Differences in composition Section III. 



b. Through changes taking place in the course of experiment : 



(1) In area Section V. 



(2) In composition, including change in power of retaining 



water when dried Section III. 



iii. Errors affecting the interpretation of the observed increase, due to : 



a. The heterogeneous composition of the increase — starch, 



sugars, proteids, oils, etc Section III. 



b. The occurrence of respiration and translocation. 



Section III. — On Changes of Composition during Insolation. 



To ascertain whetlier any of the observed increase of dry weight during 

 insolation is due to fixation of water, the gain of carbon per unit of area was 

 determined in the same half-leaves in which the increase of dry weight had 

 lieen carefully measured. For this purpose the " experimental " and 

 " control " i)ieces of leaf were analysed by combustion, and the gain in 

 carl)on content per unit of area determined by the difference. 



1'he analyses were carried out in the Caius College Chemical Labox'atory, 

 and my tlianks are due t(j Dr. Ruheraann, of Caius College, for placing his 

 ai)])aratuH at my disposal, and giving me every facility. 



If photosynthesis resulted in the accumulation of some known single 

 suV)8tance, the increase of dry weight could be accurately calculated from 

 the gain of carbon. As, however, starch, various sugars, proteids, and even 

 oils may 1)0 jn-oduccul, any basis for calculation must be in some degree 

 arl)itrary, aiul j)erfect correspondence between the calculated increase and 

 that actually observed is not to be expected. 



There are, l)e8ides, other minor sources of discrepancy. Tiie apparent 

 gain of dry w(M'g]it which would result from any shrinkage ui area of the 

 experim(!iit,al lialf-leaf would have the composition of the general leaf 



